There is a fundamental gap in our understanding of neural network development in young children who stutter near symptom onset, and developmental trajectories that are associated with persistence versus recovery. The type of coordinated neuronal activity that is necessary for fluent speech depends upon inter- connections within and among large-scale neural networks, and these critical connections develop during childhood. Our long-term goal is to find empirically-based, early neural markers for persistent stuttering, findings that may eventually inform the clinical diagnosis and treatment of childhood stuttering. The overall objective of the present application is to find neural network based mechanisms that are associated with persistent stuttering. Guided by data from the current project period and from a neurocomputational model of speech sequencing, our central hypothesis is that stuttering emerges and persists due to aberrant functional and structural organization within and between major brain networks, specifically those involving left cortico-basal ganglia motor circuits (?cortico-BG network?) and their interactions with other intrinsic connectivity networks (ICNs) during development. The rationale for the proposed research is that by empirically studying how maturation of neural networks of children who stutter (CWS) differ from that of their normally fluent peers, a better understanding of the complex etiology and mechanisms underlying persistent childhood stuttering should result. Guided by strong preliminary data, the central hypotheses will be tested by pursuing two specific aims: 1. Delineate and compare neurodevelopmental trajectories of the left cortico-BG network in children who stutter and typically speaking children. 2. Identify how neurodevelopmental trajectories of the left cortico-BG network and ICNs relate to both childhood stuttering as well as the risk for such stuttering to persist. The proposed work is innovative, as it will be the first series of studies designed specifically to characterize whole brain network anomalies specific to CWS, which may serve as a highly predictive neural marker for persistent stuttering during early childhood. Findings will be significant, because the expected results will elucidate, for the first time, whether major neural networks connect, organize, and mature differently in CWS and whether such differences are associated with the perpetuation of stuttering symptoms during development. Such results will have an important positive impact, by identifying neural network markers apt to predict eventual persistence versus recovery during early phases of symptom progression as well as provide developmentally- appropriate areas to consider in preventive and therapeutic interventions for childhood stuttering.